AI Voice and Deepfake Scams: When a Familiar Voice Isn't Real

For years, the advice was simple: if you recognize the voice, you can trust the call. That rule no longer holds. With a few seconds of audio, scammers can now clone a person’s voice, and with a little more, they can fake video too. A panicked call that sounds exactly like your grandchild, or a video that looks like a trusted official, may be entirely manufactured.

This is the newest layer on top of older scams, and it makes them far more convincing. This guide explains how AI voice and deepfake scams work, shows a real example, and gives you ways to protect your family that still work even when your ears and eyes are fooled.

What it is

An AI voice or deepfake scam uses artificial intelligence to imitate a real person, by cloning their voice or faking their face on video, to make a scam more believable. It is usually a tool layered onto a familiar scam, such as the grandparent emergency, a romance con, a fake executive request, or an investment pitch from a “celebrity.”

What makes it dangerous is that it attacks the very signals we use to trust, a loved one’s voice or a familiar face, and turns them into weapons.

How it works

1. A scammer collects a short sample of someone’s voice or image, often from social media, a voicemail, or a video.
2. They use AI tools to clone the voice or create a realistic fake video.
3. They place a call or send a message as that trusted person, usually with an urgent request for money or information.
4. Because the voice or face seems real, the target acts before thinking to verify.

The same technology powers fake celebrity endorsements for investment and crypto schemes, and “boss” calls that pressure an employee to send a wire. Only a few seconds of audio can be enough to produce a convincing clone.

A real example

The Nguyen family gets a call late at night. Their daughter’s voice, crying, says she has been in an accident and needs bail money fast, and begs them not to tell anyone. The voice has her exact tone and speech patterns. A “lawyer” takes over with instructions to wire money immediately. The parents are moments from sending it when the father texts their daughter directly and she replies, confused, from her apartment. The voice on the call had been cloned from a short video she had posted online. Nothing about it sounded fake, which is the entire danger.

By the numbers

• A scammer needs only a few seconds of audio to clone a familiar voice (FTC).
• The FBI’s IC3 received more than 3,100 complaints from older adults referencing AI in 2025, with losses topping $350 million (FBI IC3).
• The FBI has warned that criminals increasingly use generative AI to make fraud more believable and harder to spot (FBI IC3).

Red flags to watch for

• An urgent call or video from a loved one or official asking for money or secrecy.
• Pressure to act immediately, before you can verify.
• A request for payment by gift card, wire, or cryptocurrency.
• A “celebrity” or “executive” endorsing an investment or ordering a transfer.
• Small glitches in a video call, or a caller who avoids switching to a live, two-way video.

How to protect yourself

1. Treat a familiar voice as no longer being proof. Verify before you act, every time.
2. Hang up and call the person back on a number you already have.
3. Set a family password, a simple word only your family knows, to confirm identity in an emergency. It works even against a perfect voice clone.
4. Ask a question only the real person could answer, and be wary of anyone who resists.
5. Limit the raw material. Scammers pull voice and video from public social media and details from data-broker sites. Tightening privacy settings, and removing your information from broker sites (a privacy or data-removal service can do this), gives them less to work with.
6. Talk about AI scams with your family now, so an unbelievable-sounding call is met with a plan instead of panic.

If you’ve already responded

Contact your bank, card company, or payment app right away to try to stop or reverse it. Save the messages and any recordings, and report it to the FBI at ic3.gov and the FTC at ReportFraud.ftc.gov. It is not your fault; these fakes are designed to fool careful, loving people.

In the news

• FBI report: internet crime losses hit a record in 2025 (AARP)
• FBI IC3: Criminals use generative AI to facilitate financial fraud

Sources

• FTC: Scammers use AI to enhance their family emergency schemes
• FBI IC3: Generative AI and financial fraud
• FBI: Scammers Target Older Adult Victims
• FBI IC3: Elder Fraud